The role of matrix vesicles (MV) in hard tissue mineralization has been debated since these enigmatic particles were first observed more than 40 years ago in the matrix of calcifying cartilage. In those studies, and many more recent investigations, it has been shown that MV are present in the matrix of many mineralizing tissues, including cartilage, bone and dentin, and that they are often seen as the site where mineral crystals first form. When these tissues mineralize, however, the vast majority of the mineral is found associated with collagen fibrils, which are completely permeated and encased in mineral. A key question in the proposed mechanism of MV initiated matrix mineralization is to explain how mineral crystals formed within MV initiate collagen calcification. This is the problem at which this proposal is aimed. To answer this question, we propose three hypothetical models which each provide a possible answer to the question. The first model focuses on the role of MV enzymes in altering the relative concentration of inorganic phosphate (Pi) and inorganic pyrophosphate (PPi) to make the extracellular milieu permissive for mineral to nucleate on matriix molecules directly. The second model predicts that MV nucleate mineral internally, release small crystals and that these crystals seed the mineralization of the matrix. The third model specificies that MV bind to collagen fibrils at specific sites, and that while bound, they facilitate crystal entry into the colagen fibrils. Three Specific Aims are proposed to test these models and distinguish amongst them. The first experiments are designed to decide whether MV must come into proximity with collagen fibrils to correctly mineralize them. If proximity is required, then the first model in unlikely to be correct. The next experiments will examine the proposition that MV binding is necessary for collagen mineralization. These experiment will allow us to distinguish between the 2nd and 3rd models. Another set of experiments will examine the possibility that MV associated enzymes remodel the collagen fibrils to facilitate crystal entry. Depending on whether the remodeling enzymes are MV associated or are free in the matrix, these data will support either model 2 or model 3. Finally, we will construct synthetic models of MV, and test the function of MV proteins and lipids in isolation and together to better distinguish among the proposed hypotheses. When these experiments are completed, our knowledge of how MV initiate matrix calcification will be greatly enhanced.